Oral GFH375 posted responses in a 74-patient KRAS G12D phase 1
A Nature Medicine phase 1 paper dated 15 September 2026 gave oral GFH375 to 74 people with KRAS G12D tumours who had already had a lot of treatment. The 22 people in the climb cohort had no dose-stopping toxicity; the phase 2 suggestion was 600 mg once a day. In radar pancreatic radar cancer radar the radar confirmed radar objective radar response radar rate was 35.1 percent (13/37). New Scientist and EurekAlert pages checked on 15 September did not carry their own matching trial write-up.
Science··Night
74 patients, no dose-limiting toxicity in 22
The Nature Medicine text is dated 15 September 2026. GFH375 is an oral candidate that binds KRAS G12D in both radar the radar GTP radar-bound on and GDP-bound off states. Primary endpoints were safety, the highest dose that can be climbed, and the dose suggested for phase 2. Of 74 patients, pancreatic ductal adenocarcinoma accounted for 43, with 16 lung tumours of the non-small-cell type and 10 colorectal tumours. The 22 people in the dose-climb cohort had no dose-stopping toxicity recorded; the amount suggested for phase 2 was 600 mg once a day. The maximum tolerated dose was not found, the paper says. The card does not treat this as an approved drug claim.[1]
Pancreatic response and grade 5
One grade 5 event tied to treatment is written as radar septic radar shock (1.4 percent). Grade-3 or worse treatment-related events were 36.5 percent, 27/74. In the pancreatic group the confirmed objective response was radar 35.1 percent, 13/37, with median response length 5.6 months and median progression-free survival 5.5 months. A median overall-survival row continues in the paper; this card does not copy that row in full. It is an early-phase safety and signal text. There is no control arm.[1]
One trial paper
The 74-patient phase 1, the 600 mg suggestion and the 35.1 percent pancreatic rate sit in the same Nature Medicine paper. New Scientist and EurekAlert pages and feeds on 15 September did not carry their own matching trial write-up. The card does not treat 35.1 percent as a market launch or a standard-of-care claim. The figure is a confirmed response rate. The second-publisher check was empty.[1]
